Accident Detection and Reporting System Using IOT

  • C. Santhanakrishnan, Lin Sanjo, Aswin Jinachandra


During the most recent decade it is very much seen that Computer vision has supplanted human vision in the field of prepared to catch a scope of reasonable peculiarities. during this paper, we will in general propose to be told peculiarities by misusing each conventional and unusual video and afterward structure a framework that can report Exigencies to the close by Emergency Service Units. Deciphering the variation from the norm in instructing recordings fantastically time overpowering, so to keep away from these challenges we will in general adapt abnormality through the profound different example positioning system by supporting weak labelled training recordings, region unit at video level as opposed to cut level. conventional and irregular recordings zone unit contemplated as stuff and video fragments as occurrences in Multiple Instance Learning (MIL) and in this way anticipating an Anomalous score dependent on which an edge will be created whereupon the framework distinguishes whether a mishap has happened or not. An enormous scope informational collection which contain 100 hrs of recordings is presented. To start with, general peculiarity location considering every single unusual fragment are gathered as the whole gang ordinary exercises as another. Second, for dissecting every one of odd street exercises. Our exploratory Observations show that MIL strategy for bizarre street mishap location accomplishes an unusual enhancement for execution.              

How to Cite
C. Santhanakrishnan, Lin Sanjo, Aswin Jinachandra. (2020). Accident Detection and Reporting System Using IOT . International Journal of Advanced Science and Technology, 29(06), 95 - 102. Retrieved from